Time-ResolvedFluoroimmunoassayof Human ... - Clinical Chemistry

0 downloads 0 Views 931KB Size Report
Jun 21, 1982 - use in the diagnosis of gynecological emergen- cies; and the .... reacted with anti-LHa for 20 h at 0 to 4#{176}C.Passage through a Sephadex.

CLIN.CHEM.29/1, 60-64

(1983)

Time-ResolvedFluoroimmunoassay of HumanChoriogonadotropin Kim Pettersson,1 Harri Siitari,1 Ilkka Hemmil#{228},’ Erkki Soini,1 Timo L#{246}vgren,’4 Veli H#{228}nninen,2 Pirjo Tanner,2 and Uif-H#{225}kan Stenman3 We describe time-resolved fluoroimmunoassay for human choriogonadotropin involving monoclonal antibodies directed against the /3- and a-subunits. The latter antibody was labeled with europium, which was measured by counting for 1 s after the immunoreaction was completed. In the solidphase sandwich assay, both a one-step and two-step procedure were used; the respective measuring ranges were 0.7135 and 0.7-350 mt. units/L, the latter covering a 500-fold dynamic range. The CV within the assay range was between 4 and 8%, depending on the dose. Cross reactivity with lutropin in the one- and two-step procedures was 1.6% and 1.0%, respectively.

Additional Keyphrases: monoclonal antibodies ftjfr.. pin peptide hormones europium label “sandwich” assay .

#{149}

.

choriogonadotropin (hCG) of placental origin is marker for pregnancy detection (1-3). Immunological pregnancy tests have long suffered from poor sensitivity, mainly because of the inability of anti-hCG antiserum to distinguish between hCG and lutropin (luteinizing Human

the primary

hormone). With the development of antisera directed towards hCG-specific sites on the hormone’s /3-subunit, one could reliably detect hCG by RLA as early as eight days after ovulation (4). Modifications of hCG RIA (5) are now finding increasing use in the diagnosis of gynecological emergencies; and the quantitative measurement of hCG in serum is important in monitoring therapy of trophoblastic disease and other hCG-producing tumors (1, 3). Although RIA methods generally have satisfactory sensitivity, many researchers and routine laboratories have begun to search for alternative markers to radioactively labeled substances. Apart from the radiation hazards of the analyte itself to the personnel who handle radioactive substances, the iodination procedure involves conditions that may alter the antigenic properties of the tracer (5-6). We have developed a highly sensitive and specific two-site “sandwich” fluoroimmunoassay for quantiQying hCG in serum. Purified monoclonal antibodies to the hCG /3-subunit, immobilized to polystyrene tubes, are used as the solid phase. For a tracer we used purified monoclonal antibodies to the a-subunit of lutropin, to which we covalently attached

europium complexed resolved fluorometry label, we minimized

‘Wallac Biochemical Finland.

to an EDTA-denvative. By using time(TR-FLA) (7) to detect the europium background fluorescence and obtained

Laboratory,

P.O. Box 10, 20101 Turku 10,

Laboratories Ltd., Kauniainen, Finland. 3Department of Obstetrics and Gynaecology, University 2Medjx

Central Hospital, Helsinki, Finland. Address correspondence to this author. #{176}Nonstandard abbreviations: hCG, human choriogonodatropin; hCGa, a-subunit of hCG; hCG-13, 13-subunit of hCG; LH, a-subunit of lutropin; TR-FIA, time-resolved fluoroimmunoassay; IRP, International Reference Preparation (Organon); IS, International Standard (Boehringer). Received June 21, 1982; accepted Oct. 13, 1982.

60 CLINICAL CHEMISTRY, Vol. 29, No. 1, 1983

maximum

sensitivity.

hCG at about 1

With

mt. unitlL

this procedure

we could detect

in a 1-h assay.

Materials and Methods Materials Monoclonal mouse antibodies, anti-hCGf3 (lot no. 5008) and anti-LHa (lot no. 0301) purified from ascites fluid, were obtained from Medix Laboratories, Helsinki, Finland. According to the manufacturer, the cross reactions of the antihCG/3 as determined by conventional competitive RIA were as follows: hCG 100%; hCGJ3 118%; hCGa 2.2%; lutropin 0.4%; thyrotropin 0.05%. This antibody has been used in a conventional RIA and gives results comparable with those obtained with conventional 13-specific antibodies (8). Highly purified hCG standards were obtained from Organon, The Netherlands Especific activity 10 900 i. units/mg according to the First International Reference Preparation for immunoassays (IRP)1, and from Boehringer, Mannheim, F.R.G. Fspecific activity 13 500 mt. units/mg in terms of the 2nd International Standard for bioassays (IS)]. Human lutropin (specific activity 6700 mt. units/mg according to IS) as well as purified a- and /3-subunits of hCG were obtained from Boehringer. Diethylenetriaminepentaacetic acid and EDTA were obtained from Merck AG, Darmstadt, F.R.G. Bovine serum albumin and bovine globulin were purchased from Sigma Chemical Co., St. Louis, MO 63178. Eu203 was obtained from Fluka AG, Buchs, Switzerland. Other chemicals used

were of analytical-purity grade. The chloride salt of Eu was prepared by dissolving Eu203 in hydrochloric acid and evaporating the excess acid. Diazophenyl-EDTA was synthesized by a modification of the method of Sundberg et al. (9) by adding Eu before conjugating it to the protein. 2-Naphthoyltrifluoroacetone was synthesized according to the method of Reid and Calvin (10). Buffers used in the TR-FIA were: Tris HC1 (50 mmol!L, pH 7.7 at 25 #{176}C), containing 9 g of NaCl and 0.5 g of NaN3 per litre (A buffer); A buffer plus bovine serum albumin, 5 g/ L (B buffer); B buffer plus 0.5 g of bovine globulin and 1 mL of Tween 20 surfactant per litre (C buffer). Assay tubes were washed with a saline solution containing 9 g of NaCl and 0.5 g of NaN3 per litre.

Methods Conjugation of diazophe7zyl-EDTA-Eu to anti -LHa. Diazophenyl-EDTA-Eu at 20- to 40-fold molar excess was reacted with anti-LHa for 20 h at 0 to 4#{176}C. Passage through

a Sephadex G-50 column (1.5 x 20 cm) separated the conjugated protein fraction from excess reagent: each 2-mL fraction was monitored for absorbance at 280 nm and for fluorescence (see below). Incorporation of diazophenylEDTA-Eu into the tyrosine and histidine residues of the protein varied from 2.5 to 10 molecules of Eu per molecule of

!gG. An incorporation ratio of four or five molecules of Eu per molecule of IgG was found to be optimal and had a negligible effect on affinity. Bovine serum albumin was added to pooled fractions to give a concentration of 5 g/L, and the pooled fractions were stored at 0 to 4 #{176}C.

Coating of polystyrene tubes with anti -hCGf3. Anti-hCG/3 antibody was immobilized by adsorption to the interior surface of 12 x 50 mm polystyrene tubes (Vapex, Calamandrana, Italy). The tubes were coated with 0.25 mL of purified antibody (10 .tg/mL) in 0.25 mL of 0.1 mol/L sodium carbonate buffer, pH 9.5, for 20 h at room temperature. After coating, we added 1 mL of B buffer to each tube and stored them at 0 to +4#{176}C. Before use, the coated tubes were washed twice with the saline solution. Time-resolved fluoroimmunoassay. This was done accordmg to two different procedures, a one-step and a two-step procedure. One-step TR-FIA: To washed, coated tubes we added 100L serum samples or hCG-standards in C buffer, followed by 100 pL of C buffer or control serum from males (hCG-free). Then 30-SO ng of Eu-labeled anti-hCGa in 50 L of C buffer, with 100 and 500 mol of diethylenetriaminepentaacetic acid and Ca, respectively, per litre, was added to each tube shortly before use to remove any dissociated free Eu. After gentle vortex-mixing, the tubes were incubated for 1 h at room temperature. We then aspirated the tubes’ contents and washed the tubes with three 1-mL portions of saline solution. Determinations were done in duplicate if not otherwise stated. Two-step TR-FIA: To washed, coated tubes we added 100pL serum samples or hCG standards in C buffer followed by 100 zL of C buffer or male control serum and then an additional 50 .tL of C buffer. After gentle vortex-mixing and incubation for 1 h at room temperature, we aspirated the tubes’ contents and washed the tubes with 1 mL of saline solution. We then added to each tube 30-50 ng of Eu-labeled anti-hCGa in 250 L of C buffer containing 20 mol ot diethylenetriammnepentaacetic acid and 100 mol of Ca per

litre. After

and allowing the tubes to stand for hour at room temperature, we aspirated their

vortex-mixing

an additional

contents, then washed saline solution.

the tubes with three 1-mL portions of

Measurement of fluorescence. After washing the assay tubes, we measured the Eu bound to the solid phase as a 2naphthoyltrifiuoroacetone chelate by using a single-photon counting time-resolved fluorometer as described by Soim and Kojola (11). (Details of the formation of the 2-naphthoyltrifluoroacetone chelate are to be published elsewhere.) This fluorometer has a xenon 1000-Hz flash lamp. The measuring time was 1 s, the delay time 400 j.ts, and the counting time 500 .cs. The excitation wavelength was 340 nm and the emission wavelength 613 nm. The background fluorescence of 250 cps for noncoated polystyrene tubes was subtracted from all readings.

Results Sensitivity and Specificity ofthe Assay The dose-response curve obtained by TR-FIA is fairly linear, and the fluorescence intensity is nearly directly proportional to the antigen concentration over a large concentration range (Figure 1). The fluorescence intensity of the zero sample (male serum) has not been subtracted in the standard curves shown. If this is done, the standard curves for both assays are linear over 80-90% of the measuring range and the signal is proportional to the dose. By the one-step procedure, a measuring range of 0.05-10 ng/mL for the Boehringer standard (2nd IS) was obtained. This corresponds to 0.7-135 mt. units/L, there being a 200-

fold difference between the highest and lowest measurable concentrations. In the two-step procedure, the measuring range was even larger, 0.05-25 ng/mL, or 0.7-350 mt. units/ L-a 500-fold difference. The minimum detection limit of the one-step procedure was better than 0.7 mt. unitfL, but this was the lowest standard dilution used in the assay. If the minimum detection limit is defined as the lowest

10

A

B

0

z

hCG

0 U

hCG (Boehringer)

ci) z 0 C.) w C.)

hCG,l

z Iii

U C,) U

0 -I U.

f...

io2 0

0.1

1

10

100

1000

0

0.1

1

10

100

1000

ng/mL

Fig. 1. One-step (A) and two-step TR-FIA ( results, showing the standard curves with the Boehringer (2nd IS) and the Organon (1st IRP) hCG standards The cross reactions of anti-hCG antibodies with lutropin (hLH) and tree a-hCG subunits are also shown. Each point on the standard curves representthe mean of triplicates; bars indicate I SD. Points on the lutropin and a-subunit cross-reactioncurves represent the mean of duplicates. The dashed horizontal line denotes the background level plus 3 SD

CLINICALCHEMISTRY, Vol. 29, No. 1, 1983

61

concentration of analyte that produces fluorescence that differs significantly from that produced by the zero sample, then the detection limit is about 0.1 unitfL. This value was defined as the hCG concentration corresponding to the mean fluorescence of 10 aliquots of the zero sample plus 3 SD. For the one-step procedure the background fluorescence was 166 (SD 19) cps, and for the two-step assay it was 428 (SD 58)

cpa. A precision proffle calculated from 10 replicates of each hCG concentration assayed according to the two-step PHYLA is shown in Figure 2. When samples are run in duplicate and a counting time of 1 s or 10 s is used, the working range of the assay is from 0.06 ng/mL, or 0.9 mt. unit/L, to at least 37 ng/mL, or 500 units/L, respectively, if a CV of 10% is taken as the discrimination limit.

Table 1. SensitIvities of One-Step and Two-Step TR-FIA of hCG with Two Different Standard Preparations Sensitivity, pg/L (and Standard preparation Boehrtnger

Normal sore

0.05 (0.7)

0.05 (0.7)

0.05 (0.7)

0.06 (0.8)

0.1 (1.1)

0.13(1.4)

0.2 (2.2)

0.24 (2.6)

(n

=

15)

standard

One-step method Two-step method Organon standard One-step method Two-step

method

‘Background

procedure. background

40

mt. unlts/L)’

Control sara (n = 10)

plus 3 SD for no. of determinations and specimens indicated.

The ratio between the maximum signal and the was 265 in the one-step assay and 210 in the

two-step assay. The cross reaction of anti-hCGa antibody with lutropin was 1.0% for the two-step assay and 1.6% for the one-step assay. These figures represent the maximum cross reaction at hCG concentrations of 1-10 mt. units/L. At higher concentrations the cross reactions were lower because of slight non-parallelism between the respective hCG and lutropin dose-response curves. In the one-step assay the curves were parallel up to a lutropin concentration of 100 ng/mL, but the maximum signal obtained with lutropin was only about 10% of that obtained with hCG. In the two-step assay lutropin gave a less steep dose-response curve than hCG. The cross-reaction figures for the Organon standard were about four times higher. The cross reaction with free asubunit was lower than with lutropin.

30

20

Influence of Free Subunits

!-

10

The effect of free subunits

on the assay

was

studied

by

adding a one- to fourfold excess (on a weight basis) of hCGa

I

I

0.037

I

0.11

0.33

I

I

I

I

1

3

9

27

10

I

II

ng/mL Fig. 2. Precision

profile of TR-FIA of hCG

determinationsfor each hCG concentration (2nd IS, Boehringer) accordingto the two-step procedure. The CV (in %) for each dose was calculated for counting times of 1 s (0) and lOs (S). The lowest curve (#{149}) represents the precision profile calculatedfrom duplicate measurements Mean of 10

0

z

0 U U

U) U) I-.

The mean background

for 15 individual serum samples from normal healthy men was 158 (SD 38) cps for the one-step procedure and 470 (SD 65) cps for the two-step procedure. The corresponding sensitivities measured as 3 x SD in these normal samples from men were less than 0.05 ng/mL (2nd IS) for the one-step procedure and 0.06 for the two-step procedure (Table 1). The values obtained with the Organon standard (1st IRP) were about fourfold those obtained with the Boehringer standard (2nd IS) given above, but the shape of the standard curves and the maximum fluorescence observed were practically identical. measured

The shapes of the standard curves differ for the two assay methods. As expected, the standard curve for the one-step procedure reaches a maximum, after which it decreases with increasing hCG concentration. This reflects the situa-

tion in which the tracer increasingly binds to an excess of antigen, which itself cannot bind to the saturated solidphase antibody. In the two-step assay, the fluorescence intensity reached a plateau that was approximately twice as high as the maximum signal obtained in the one-step 62

CLINICAL

CHEMISTRY, Vol. 29, No. 1, 1983

z

10

4

0

U U

U

z U

U

U) U

0 -i U.

I..

0

I

0.5

I

I

I

I

2.5

10

40

160

ng/mL

Fig.3. Influence

of free a-subunit on the hCG standard curve Different concentrations of a hCG-standard (Boehringer)without (V) or in the presence of one (0), two (0), and four (A) times as much a-subunits as hCG were assayed by the one-step TR-FIA procedure

to standards containing known amounts of hCG. In the onestep assay a one- to twofold excess of free a-subunits had a very small effect on the lower part of the standard curve (Figure 3). However, the measurement range for high concentrations was significantly decreased. With a fourfold excess of free a-subunits, the measurement range was severely limited and the response curve was displaced to the right (Figure 3). In the two-step assay, no interference was observed even at a fourfold excess of free a-subunits. This corresponds to a 12-fold molar excess. The effect of free /3-subunits was fairly small. The presence of /3-subunits in a concentration similar to that of hCG caused a measurable interference only at the uppermost part of the standard curve. The effect was similar in both assays (Table 2). The effect of higher concentrations of /3subunits was not studied because they are not clinically relevant

(12).

Table 2. Effect on the One-Step and Two-Step TRFIA Assays of Increasing the hCGj3 Subunits Fluorescence,

hCG/ “g/L

cpa x iO

One-step

Two-step

44.5

67.0 65.0 67.5 64.4

45.3

0.5 1.25 5 25

43.9 44.9

41.4

100 250

26:4

hCG concn, 25 aQIL

Correlation between the One-Step and Two-Step Assays Figure 4 shows the correlation between the two fluoroimmunoassay procedures. In the one-step assay the samples were usually tested in two different dilutions to ascertain

that the concentration of hCG in the sample was within the measuring range. Doing this resulted in very good agreement (r = 0.98). Results were slightly lower with the onestep procedure.

4

a’

a very high sensitivity for a 1- or 2-h assay. The observed specificity of the present hCG TR-FIA, mainly determined by the specificity of the anti-hCGf3 antibody used, is very similar to that obtained by using the same anti-hCGf3 antibody in a conventional RIA (8). Although the cross reaction with LH is only 1.0-1.6%, this has to be considered if the maximum sensitivity of the assay is utilized, e.g., even 10 mt. units of lutropin per litre would be detected in the one-step assay and therefore it would be an

advantage

to dilute the samples

five- or 10-fold in routine

assays. Diluting would also extend the assay range up to 1350 mt. units/L while retaining a sensitivity of about lint. unitiL in the one-step assay. Theoretically, an immunoassay of the present type should

free subunits,

but a clear dose-response

was

observed with hCGa. This response was less than 1% of that observed with hCG and was possibly caused by the contamination of the hCGa preparation with hCG. Addition of free hCG subunits in concentrations occurring in serum samples (12-14) had a negligible effect on the results. The two-step method was especially insensitive to subunit excess. The two hCG standards that we used differed in potency by a factor of four-not surprising, because they had been calibrated against different standards. The 1st IRP has been recommended for use in immunoassay but, in practice, the 2nd IS is much more widely used. Sensitivity and specificity calculations based on the 2nd IS are therefore comparable with those generally used in clinical practice. A prominent feature of the hCG TR-FIA is the very wide

radiolabeled

instead of Eu-labeled anti-hCGa was used, this conclusion. The choice between the one-step and two-step assay methods depends on their application. The one-step method is more rapid, but because of the shape of the standard curve it is necessary to assay at least two dilutions of the sample, and therefore the two-step method seems to have practical

‘.

3

seem

10

n-38 Y-O.96o-55 rO.985

#{149} 0

.

.

#{149}0

10

demonstrate that very sensitive and rapid can be developed with time-resolved fluorescence as the detection method. The combination of speed, sensitivity, and specificity obtained in this hCG TR-FIA is probably unequaled at present. The properties of this fluoroimmunoassay method are the result of several factors. The newly developed TR-FIA technology (11) allows very sensitive detection of the tracer. By using purified monoclonal antibodies for both coating the solid phase and for the tracer, very high concentrations of antibody can be added to the sample tubes. This means that sufficient antibody-antigen binding is quickly attained. This explains why we obtained

measurement range. This is a result of the sensitive fluorescence detection method and the high concentrations of antibody used. In a two-site sandwich assay of the type used here, the upper part of the measurement range is mainly limited by the amounts of antibodies added to the sample tubes. However, the present assay does not differ in principle from other sandwich assays (15) except for the type of label used. This suggests that the high sensitivity at least partly results from the use of time-resolved fluorescence. Preliminary results obtained by a similar assay, in which

Ia’

C

Our results

immunoassays

not measure

33.5

Cu

Discussion

0

2

..

io

I0

Two-slep

1o

TR-FA

hCG. int.units/

Fig. 4. Comparisonof and two-step TR-FIA

38 serum

I.

10

to confirm

advantages. This is especially true for hCG, the concentration of which varies over such a wide range. For many other hormones, such as lutropin, the measurement range of the assay would cover the physiological concentration range. Engvall et al. (16) recently described a very sensitive enzyme-labeled immunosorbent sandwich assay for alpha-

hCG valuesobtained bytheone-step fetoprotein involving two monoclonal antibodies reacting with different sites on the antigen. This assay is basically CLINICAL CHEMISTRY, Vol. 29, No. 1, 1983 63

similar to the present TR-FIA method, the main difference being the use of an enzyme instead of Eu for labeling the second antibody. Many of the advantages observed in the fluoroimmunoassay described here were also achieved in the enzyme immunoassay: a sensitivity better than 1 p.g/L in a 2-h assay and a linear dose response over a 100-fold concentration range. The TR-FIA method has certain advantages when compared with this and other enzyme immunoassays. The dynamic range of the detection method is wider in TR-FIA; in this study the signal/noise ratio was >200. This can probably not be achieved by routine photometric methods. Even when compared with radioisotopic methods, the measurement range obtained by TR-FLA seems to be superior. In TR-FIA very good counting precision is obtained, even with a counting time of only 1 s. The main advantages of enzyme immunoassays are shared by our TR-FIA method. The reagents are nonhazardous and stable and the use of the assay is not limited by radiation regulations. This is especially important for emergency applications such as early detection of pregnancy (5, 8). These advantages, in combination with better sensitivity and shorter assay time, make time-resolved fluoroimmunoassay a very competitive immunoassay method.

References 1. Marrs R, Mishell D. Placental trophic hormones. Clin Obstet Gynecol 23, 721-737 (1980). 2. Frances R, Batzer MD. Hormonal evaluation of early pregnancy. Fertil Steril 34, 1-13 (1980). 3. Gi-iffT, Ross MD. Clinical relevance of research on the structure of human chorionic gonadotropin. Am J Obstet Gynecol 129, 795808 (1977). 4. Mishell DR, Nakamura RM, Barberia JM, Thorneycroft lB.

64

CLINICAL CHEMISTRY, Vol. 29, No. 1, 1983

detection of human chorionic gonadotropin in serum in normal human gestation. Am J Obstet Gynecol 118,990-991 (1974). 5. Seppala M, Tontti K, Ranta T, et al. Use of a rapid hCG-betasubunit radioimmunoasaay in acute gynaecological emergencies. Lancet i, 165-166 (1980). 6. Groom GV. The measurement of human gonadotrophins by radioimmunoassay. J Reprod Fertil 51, 273-286 (1977). 7. Soini E, HemmilA I. Fluorounmunoassay: Present status and key problems. Clin Chem 25, 353-361 (1979). Initial

8. Stenman U-H, Tanner P, Ranta T, et al. Monoclonal antibodies to chorionic gonadotropins. Use in a rapid radioimmunoassay for gynecologic emergencies. Obstet Gynecol 59, 375-377 (1982). 9. Sundberg MW, Meares CF, Goodwin DA, Diamanti CL Selective binding of metal ions to macromolecules using bifunctional analogs of EDTA. J Med Chem 17, 1304-1307 (1974). 10. Reid JO, Calvin M. Some new -diketones containing the trifluoromethyl group. JAm Chem Soc 72, 2948-2952 (1950). 11 Soini E, Kojola H. Time-resolved fluorometry of lanthanide chelates-a new generation of non-isotopic immunoassays. Clin Chem this issue (1983). 12. Vaitukaitis JL Changing placental concentrations of human chorionic gonadotropin and its subunits during gestation. J Clin Endocrinol Metab 38, 755 (1974). 13. Hagen C, MeNeilly AS. The gonadotropic hormones and their subunits in human maternal and fetal circulation at delivery. Am J Obstet Gynecol 121, 926-930 (1975). 14. Vaitukaitis JL, Ebersole ER Evidence for altered synthesis of human choriomc gonadotropin in gestational trophoblastic tumors. J Clin Endocrinol Metab 42, 1048 (1976). 15. Sekiya T, Furuhashi Y, Goto S, et al. Specific enzyme immunoassay for human chorionic gonadotrophin. Acta Endocrinol 97,562568 (1981). 16. Uotila M, Ruoslahti E, Engvall E. Two-site enzyme immunoaasay with monoclonal antibodies to alpha-fetoprot.ein. J Immunol Methods 42, 11-15 (1981).